INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

construct

-0.68

advertisement

-0.66

 Sector

-0.66

punk

-0.65

Pg

-0.63

pee

-0.59

 races

-0.58

ajor

-0.58

agra

-0.57

oples

-0.57

POSITIVE LOGITS

éĹĺ

0.79

åŃĲ

0.74

andise

0.71

irl

0.66

bard

0.65

enture

0.64

 commencement

0.63

sted

0.63

Tid

0.61

 weave

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.